dravoice 0.1.2 → 0.1.4

package/src/v2/inspect.js

@@ -1,67 +1,67 @@
-export function renderInspectV2(profile) {
-  const lines = [
-    "Dravoice V2 Profile",
-    "",
-    `Source: ${profile.source.documentCount} document(s), ${profile.source.wordCount} words, ${profile.source.sentenceCount} sentences`,
-    `Corpus confidence: ${capitalize(profile.source.confidence.band)} - ${profile.source.confidence.message}`,
-    "",
-    "Feature families:",
-  ];
-
-  for (const [name, family] of Object.entries(profile.families)) {
-    lines.push(`- ${name}: ${family.confidence}`);
-    if (family.warnings.length) {
-      for (const warning of family.warnings) {
-        lines.push(`  warning: ${warning}`);
-      }
-    }
-  }
-
-  lines.push("", "Feature details:");
-  for (const [name, family] of Object.entries(profile.families)) {
-    lines.push(`- ${name}: ${featureSummary(name, family.features)}`);
-    if (family.revisionHandles?.length) {
-      lines.push("  Revision handles:");
-      for (const handle of family.revisionHandles) {
-        lines.push(`  - ${handle}`);
-      }
-    }
-  }
-
-  lines.push("", "Guidance:");
-  for (const rule of profile.guidance.draftingRules) {
-    lines.push(`- ${rule}`);
-  }
-
-  lines.push("");
-  return `${lines.join("\n")}`;
-}
-
-function capitalize(value) {
-  return value.charAt(0).toUpperCase() + value.slice(1);
-}
-
-function featureSummary(name, features) {
-  if (name === "rhythm") {
-    return `sentenceWords.median=${features.sentenceWords.median}; paragraphWords.median=${features.paragraphWords.median}; listDensity=${features.listDensity}; quoteDensity=${features.quoteDensity}`;
-  }
-  if (name === "lexical") {
-    return `wordCount=${features.wordCount}; contentTypeTokenRatio=${features.vocabularyRichness.contentTypeTokenRatio}; wordLength.median=${features.wordLength.median}; maskedCharacterFourgrams=${features.maskedCharacterFourgrams?.length ?? 0}; functionWordBigrams=${features.functionWordBigrams?.length ?? 0}`;
-  }
-  if (name === "register") {
-    return `primary=${features.primary.value} (${features.primary.score}); alternates=${features.scores.slice(1, 4).map((score) => `${score.value}:${score.score}`).join(", ")}`;
-  }
-  if (name === "discourse") {
-    return `transitionRates=${Object.entries(features.transitionRates).map(([key, value]) => `${key}:${value}`).join(", ")}; sentenceCallbacks=${features.sentenceCallbacks}`;
-  }
-  if (name === "rhetoricalShape") {
-    return `openingMoves=${features.openingMoves.slice(0, 5).join(" -> ") || "none"}; moveTrigrams=${features.moveTrigrams?.slice(0, 2).map((item) => `${item.value}:${item.count}`).join(", ") || "none"}; commonSequences=${features.commonSequences.slice(0, 3).map((item) => `${item.value}:${item.count}`).join(", ") || "none"}`;
-  }
-  if (name === "evidence") {
-    return `evidenceSentenceRate=${features.evidenceSentenceRate}; claimSentenceRate=${features.claimSentenceRate}; supportedClaimRate=${features.supportedClaimRate}; unsupportedClaimRate=${features.unsupportedClaimRate}; evidenceTypes=${features.evidenceTypes.map((item) => `${item.value}:${item.count}`).join(", ") || "none"}`;
-  }
-  if (name === "structure") {
-    return `sectionWords.median=${features.sectionWords.median}; headingCount.median=${features.headingCount.median}; listDocumentRate=${features.listDocumentRate}; quoteDocumentRate=${features.quoteDocumentRate}`;
-  }
-  return JSON.stringify(features);
-}
+export function renderInspectV2(profile) {
+  const lines = [
+    "Dravoice V2 Profile",
+    "",
+    `Source: ${profile.source.documentCount} document(s), ${profile.source.wordCount} words, ${profile.source.sentenceCount} sentences`,
+    `Corpus confidence: ${capitalize(profile.source.confidence.band)} - ${profile.source.confidence.message}`,
+    "",
+    "Feature families:",
+  ];
+
+  for (const [name, family] of Object.entries(profile.families)) {
+    lines.push(`- ${name}: ${family.confidence}`);
+    if (family.warnings.length) {
+      for (const warning of family.warnings) {
+        lines.push(`  warning: ${warning}`);
+      }
+    }
+  }
+
+  lines.push("", "Feature details:");
+  for (const [name, family] of Object.entries(profile.families)) {
+    lines.push(`- ${name}: ${featureSummary(name, family.features)}`);
+    if (family.revisionHandles?.length) {
+      lines.push("  Revision handles:");
+      for (const handle of family.revisionHandles) {
+        lines.push(`  - ${handle}`);
+      }
+    }
+  }
+
+  lines.push("", "Guidance:");
+  for (const rule of profile.guidance.draftingRules) {
+    lines.push(`- ${rule}`);
+  }
+
+  lines.push("");
+  return `${lines.join("\n")}`;
+}
+
+function capitalize(value) {
+  return value.charAt(0).toUpperCase() + value.slice(1);
+}
+
+function featureSummary(name, features) {
+  if (name === "rhythm") {
+    return `sentenceWords.median=${features.sentenceWords.median}; paragraphWords.median=${features.paragraphWords.median}; listDensity=${features.listDensity}; quoteDensity=${features.quoteDensity}`;
+  }
+  if (name === "lexical") {
+    return `wordCount=${features.wordCount}; contentTypeTokenRatio=${features.vocabularyRichness.contentTypeTokenRatio}; wordLength.median=${features.wordLength.median}; maskedCharacterFourgrams=${features.maskedCharacterFourgrams?.length ?? 0}; functionWordBigrams=${features.functionWordBigrams?.length ?? 0}`;
+  }
+  if (name === "register") {
+    return `primary=${features.primary.value} (${features.primary.score}); mixedRegister=${features.mixedRegister ? "yes" : "no"}; markerSets=${features.markerSets?.length ?? 0}; alternates=${features.scores.slice(1, 4).map((score) => `${score.value}:${score.score}`).join(", ")}`;
+  }
+  if (name === "discourse") {
+    return `transitionRates=${Object.entries(features.transitionRates).map(([key, value]) => `${key}:${value}`).join(", ")}; sentenceCallbacks=${features.sentenceCallbacks}`;
+  }
+  if (name === "rhetoricalShape") {
+    return `openingMoves=${features.openingMoves.slice(0, 5).join(" -> ") || "none"}; moveTrigrams=${features.moveTrigrams?.slice(0, 2).map((item) => `${item.value}:${item.count}`).join(", ") || "none"}; commonSequences=${features.commonSequences.slice(0, 3).map((item) => `${item.value}:${item.count}`).join(", ") || "none"}`;
+  }
+  if (name === "evidence") {
+    return `evidenceSentenceRate=${features.evidenceSentenceRate}; claimSentenceRate=${features.claimSentenceRate}; supportedClaimRate=${features.supportedClaimRate}; unsupportedClaimRate=${features.unsupportedClaimRate}; evidenceTypes=${features.evidenceTypes.map((item) => `${item.value}:${item.count}`).join(", ") || "none"}`;
+  }
+  if (name === "structure") {
+    return `sectionWords.median=${features.sectionWords.median}; headingCount.median=${features.headingCount.median}; maxHeadingDepth.median=${features.maxHeadingDepth?.median ?? 0}; sectionOrderPatterns=${features.sectionOrderPatterns?.length ?? 0}; listDocumentRate=${features.listDocumentRate}; quoteDocumentRate=${features.quoteDocumentRate}`;
+  }
+  return JSON.stringify(features);
+}

package/src/v2/io-utils.js

@@ -0,0 +1,51 @@
+import fs from "node:fs";
+import path from "node:path";
+
+const DEFAULT_MAX_READ_BYTES = 5 * 1024 * 1024;
+
+export function readUtf8FileBounded(filePath, {
+  label = "File",
+  maxBytes = DEFAULT_MAX_READ_BYTES,
+  rejectBinary = true,
+} = {}) {
+  const resolved = path.resolve(filePath);
+  const stats = fs.statSync(resolved);
+  if (!stats.isFile()) {
+    throw new Error(`${label} is not a regular file: ${filePath}`);
+  }
+  if (stats.size > maxBytes) {
+    throw new Error(`${label} ${displayPath(filePath)} exceeds the ${maxBytes} byte limit.`);
+  }
+  const contents = fs.readFileSync(resolved, "utf8");
+  if (rejectBinary && contents.includes("\0")) {
+    throw new Error(`${label} ${displayPath(filePath)} looks like binary-looking text and cannot be analyzed.`);
+  }
+  return contents;
+}
+
+export function readJsonFileBounded(filePath, options = {}) {
+  const label = options.label ?? "JSON file";
+  const contents = readUtf8FileBounded(filePath, { ...options, label });
+  try {
+    return JSON.parse(contents);
+  } catch (error) {
+    throw new Error(`${label} ${displayPath(filePath)} is not valid JSON: ${error.message}`);
+  }
+}
+
+export function writeUtf8FileSafely(filePath, contents) {
+  rejectSymlink(filePath);
+  fs.mkdirSync(path.dirname(filePath), { recursive: true });
+  rejectSymlink(filePath);
+  fs.writeFileSync(filePath, contents, "utf8");
+}
+
+function rejectSymlink(filePath) {
+  if (fs.existsSync(filePath) && fs.lstatSync(filePath).isSymbolicLink()) {
+    throw new Error(`Refusing to write through symlink: ${filePath}`);
+  }
+}
+
+function displayPath(filePath) {
+  return String(filePath).split(path.sep).join("/");
+}